Information Diffusion Approach to Cold-Start Problem

Cold-start problem is eminent when dealing with new users and new or updated items in web environments. To approach the problem we draw motivation from the innovation diffusion theory. We thoroughly analyzed the information diffusion in the studied large scale corporate portal and discovered that innovators play the key role in the information spread. Applying these findings we introduce a novel effective solution to the cold-start problem by providing recommendations based on the browsing features of identified selected innovators. The system has been experimentally tested and evaluated.